Idea of new tool / functionnalities for YOLO image processing tools: yolosegment2json

[yvanlebras]

Hi wonder Björn, Anup, Yi!

Seems to me of interest to have a new tools (or option in the existing json2yolosegment one) so we can also convert Yolo segemnt text files into JSON one and use it as input of anylabeling GxIT to correct / ameliorate automatic annotation / detection for example.

Don't hesitate to say if this make no sense...

THANK YOU for your help!

[bgruening]

ping @anuprulez @sunyi000

[bgruening]

Jolo could also just spit out the JSON imho..or are those text files as text files useful on its own in contrast to json?

[yvanlebras]

Maybe something like this can be reuseable.. https://pypi.org/project/yolo-to-labelme/ if labelme annotaiton format is compatible with anylabeling

[sunyi000]

My opinion: I think those json files are specific to Anylabeling, so if you have existing json, they must be generated by other Anylabeling instance, and the annotation could be already corrected.

We could have yolo to anylabeling json, yolo to label me json or label me to anylabeling json etc...

those text files are used for yolo training, and can be used outside Galaxy, in other software,code and many other places too.